Logo ROOT  
Reference Guide
 
Loading...
Searching...
No Matches
InterfaceUtils.hxx
Go to the documentation of this file.
1// Author: Enrico Guiraud, Danilo Piparo CERN 02/2018
2
3/*************************************************************************
4 * Copyright (C) 1995-2018, Rene Brun and Fons Rademakers. *
5 * All rights reserved. *
6 * *
7 * For the licensing terms see $ROOTSYS/LICENSE. *
8 * For the list of contributors see $ROOTSYS/README/CREDITS. *
9 *************************************************************************/
10
11#ifndef ROOT_RDF_TINTERFACE_UTILS
12#define ROOT_RDF_TINTERFACE_UTILS
13
14#include <ROOT/RDF/RAction.hxx>
16#include <ROOT/RDF/ActionHelpers.hxx> // for BuildAction
19#include <ROOT/RDF/RDefine.hxx>
21#include <ROOT/RDF/RFilter.hxx>
22#include <ROOT/RDF/Utils.hxx>
28#include <string_view>
30#include <ROOT/TypeTraits.hxx>
31#include <RConfigure.h> // for R__HAS_ROOT7
32#include <TError.h> // gErrorIgnoreLevel
33#include <TH1.h>
34#include <TROOT.h> // IsImplicitMTEnabled
35
36#include <deque>
37#include <functional>
38#include <list>
39#include <memory>
40#include <string>
41#include <type_traits>
42#include <typeinfo>
43#include <vector>
44
45class TTree;
46namespace ROOT {
47namespace Detail {
48namespace RDF {
49class RNodeBase;
50}
51}
52namespace RDF {
53template <typename Proxied>
54class RInterface;
56} // namespace RDF
57
58} // namespace ROOT
59
60/// \cond HIDDEN_SYMBOLS
61
62namespace ROOT {
63namespace Internal {
64namespace RDF {
65using namespace ROOT::Detail::RDF;
66using namespace ROOT::RDF;
67namespace TTraits = ROOT::TypeTraits;
68
69std::string DemangleTypeIdName(const std::type_info &typeInfo);
70
71ColumnNames_t
72ConvertRegexToColumns(const ColumnNames_t &colNames, std::string_view columnNameRegexp, std::string_view callerName);
73
74/// An helper object that sets and resets gErrorIgnoreLevel via RAII.
76private:
78
79public:
82};
83
84/****** BuildAction overloads *******/
85
86// clang-format off
87/// This namespace defines types to be used for tag dispatching in RInterface.
88namespace ActionTags {
89struct Histo1D{};
90struct Histo2D{};
91struct Histo3D{};
92struct HistoND{};
93struct HistoNSparseD{};
94struct Hist{};
95struct HistWithWeight{};
96struct Graph{};
97struct GraphAsymmErrors{};
98struct Profile1D{};
99struct Profile2D{};
100struct Min{};
101struct Max{};
102struct Sum{};
103struct Mean{};
104struct Fill{};
105struct StdDev{};
106struct Display{};
107struct Snapshot{};
108struct Book{};
109}
110// clang-format on
111
112template <typename T, bool ISV6HISTO = std::is_base_of<TH1, std::decay_t<T>>::value>
113struct HistoUtils {
114 static bool HasAxisLimits(T &h)
115 {
116 auto xaxis = h.GetXaxis();
117 return !(xaxis->GetXmin() == 0. && xaxis->GetXmax() == 0.);
118 }
119};
120
121template <typename T>
122struct HistoUtils<T, false> {
123 static bool HasAxisLimits(T &) { return true; }
124};
125
126// Generic filling (covers Histo2D, HistoND, HistoNSparseD, Profile1D and Profile2D actions, with and without weights)
127template <typename... ColTypes, typename ActionTag, typename ActionResultType, typename PrevNodeType>
128std::unique_ptr<RActionBase>
129BuildAction(const ColumnNames_t &bl, const std::shared_ptr<ActionResultType> &h, const unsigned int nSlots,
130 std::shared_ptr<PrevNodeType> prevNode, ActionTag, const RColumnRegister &colRegister)
131{
133 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
134 return std::make_unique<Action_t>(Helper_t(h, nSlots), bl, std::move(prevNode), colRegister);
135}
136
137// Histo1D filling (must handle the special case of distinguishing FillHelper and BufferedFillHelper
138template <typename... ColTypes, typename PrevNodeType>
139std::unique_ptr<RActionBase>
140BuildAction(const ColumnNames_t &bl, const std::shared_ptr<::TH1D> &h, const unsigned int nSlots,
141 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Histo1D, const RColumnRegister &colRegister)
142{
144
147 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
148 return std::make_unique<Action_t>(Helper_t(h, nSlots), bl, std::move(prevNode), colRegister);
149 } else {
151 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
152 return std::make_unique<Action_t>(Helper_t(h, nSlots), bl, std::move(prevNode), colRegister);
153 }
154}
155
156// Action for Histo3D, where thread safe filling might be supported to save memory
157template <typename... ColTypes, typename ActionResultType, typename PrevNodeType>
158std::unique_ptr<RActionBase>
159BuildAction(const ColumnNames_t &bl, const std::shared_ptr<ActionResultType> &h, const unsigned int nSlots,
160 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Histo3D, const RColumnRegister &colRegister)
161{
162 if (RDFInternal::NThreadPerTH3() <= 1 || nSlots == 1) {
164 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
165 return std::make_unique<Action_t>(Helper_t(h, nSlots), bl, std::move(prevNode), colRegister);
166 } else {
168 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
169 if constexpr (sizeof...(ColTypes) > 3) {
170 h->Sumw2();
171 }
172 const auto histoSlots = std::max(nSlots / RDFInternal::NThreadPerTH3(), 1u);
173 return std::make_unique<Action_t>(Helper_t(h, histoSlots), bl, std::move(prevNode), colRegister);
174 }
175}
176
177#ifdef R__HAS_ROOT7
178// Action for RHist using RHistConcurrentFiller without weights
179template <typename... ColTypes, typename BinContentType, typename PrevNodeType>
180std::unique_ptr<RActionBase>
181BuildAction(const ColumnNames_t &columnList, const std::shared_ptr<ROOT::Experimental::RHist<BinContentType>> &h,
182 const unsigned int nSlots, std::shared_ptr<PrevNodeType> prevNode, ActionTags::Hist,
183 const RColumnRegister &colRegister)
184{
186 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
187 return std::make_unique<Action_t>(Helper_t(h, nSlots), columnList, std::move(prevNode), colRegister);
188}
189
190// Action for RHist using RHistConcurrentFiller with weights
191template <typename... ColTypes, typename BinContentType, typename PrevNodeType>
192std::unique_ptr<RActionBase>
193BuildAction(const ColumnNames_t &columnList, const std::shared_ptr<ROOT::Experimental::RHist<BinContentType>> &h,
194 const unsigned int nSlots, std::shared_ptr<PrevNodeType> prevNode, ActionTags::HistWithWeight,
195 const RColumnRegister &colRegister)
196{
198 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
199 return std::make_unique<Action_t>(Helper_t(h, nSlots), columnList, std::move(prevNode), colRegister);
200}
201#endif
202
203template <typename... ColTypes, typename PrevNodeType>
204std::unique_ptr<RActionBase>
205BuildAction(const ColumnNames_t &bl, const std::shared_ptr<TGraph> &g, const unsigned int nSlots,
206 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Graph, const RColumnRegister &colRegister)
207{
209 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
210 return std::make_unique<Action_t>(Helper_t(g, nSlots), bl, std::move(prevNode), colRegister);
211}
212
213template <typename... ColTypes, typename PrevNodeType>
214std::unique_ptr<RActionBase>
215BuildAction(const ColumnNames_t &bl, const std::shared_ptr<TGraphAsymmErrors> &g, const unsigned int nSlots,
216 std::shared_ptr<PrevNodeType> prevNode, ActionTags::GraphAsymmErrors, const RColumnRegister &colRegister)
217{
219 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
220 return std::make_unique<Action_t>(Helper_t(g, nSlots), bl, std::move(prevNode), colRegister);
221}
222
223// Min action
224template <typename ColType, typename PrevNodeType, typename ActionResultType>
225std::unique_ptr<RActionBase>
226BuildAction(const ColumnNames_t &bl, const std::shared_ptr<ActionResultType> &minV, const unsigned int nSlots,
227 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Min, const RColumnRegister &colRegister)
228{
231 return std::make_unique<Action_t>(Helper_t(minV, nSlots), bl, std::move(prevNode), colRegister);
232}
233
234// Max action
235template <typename ColType, typename PrevNodeType, typename ActionResultType>
236std::unique_ptr<RActionBase>
237BuildAction(const ColumnNames_t &bl, const std::shared_ptr<ActionResultType> &maxV, const unsigned int nSlots,
238 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Max, const RColumnRegister &colRegister)
239{
242 return std::make_unique<Action_t>(Helper_t(maxV, nSlots), bl, std::move(prevNode), colRegister);
243}
244
245// Sum action
246template <typename ColType, typename PrevNodeType, typename ActionResultType>
247std::unique_ptr<RActionBase>
248BuildAction(const ColumnNames_t &bl, const std::shared_ptr<ActionResultType> &sumV, const unsigned int nSlots,
249 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Sum, const RColumnRegister &colRegister)
250{
253 return std::make_unique<Action_t>(Helper_t(sumV, nSlots), bl, std::move(prevNode), colRegister);
254}
255
256// Mean action
257template <typename ColType, typename PrevNodeType>
258std::unique_ptr<RActionBase>
259BuildAction(const ColumnNames_t &bl, const std::shared_ptr<double> &meanV, const unsigned int nSlots,
260 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Mean, const RColumnRegister &colRegister)
261{
262 using Helper_t = MeanHelper;
264 return std::make_unique<Action_t>(Helper_t(meanV, nSlots), bl, std::move(prevNode), colRegister);
265}
266
267// Standard Deviation action
268template <typename ColType, typename PrevNodeType>
269std::unique_ptr<RActionBase>
270BuildAction(const ColumnNames_t &bl, const std::shared_ptr<double> &stdDeviationV, const unsigned int nSlots,
271 std::shared_ptr<PrevNodeType> prevNode, ActionTags::StdDev, const RColumnRegister &colRegister)
272{
273 using Helper_t = StdDevHelper;
275 return std::make_unique<Action_t>(Helper_t(stdDeviationV, nSlots), bl, prevNode, colRegister);
276}
277
278using displayHelperArgs_t = std::pair<size_t, std::shared_ptr<ROOT::RDF::RDisplay>>;
279
280// Display action
281template <typename... ColTypes, typename PrevNodeType>
282std::unique_ptr<RActionBase>
283BuildAction(const ColumnNames_t &bl, const std::shared_ptr<displayHelperArgs_t> &helperArgs, const unsigned int,
284 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Display, const RColumnRegister &colRegister)
285{
287 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
288 return std::make_unique<Action_t>(Helper_t(helperArgs->first, helperArgs->second, prevNode), bl, prevNode,
290}
291
292struct SnapshotHelperArgs {
293 std::string fFileName;
294 std::string fDirName;
295 std::string fTreeName;
296 std::vector<std::string> fOutputColNames;
298 ROOT::Detail::RDF::RLoopManager *fOutputLoopManager;
299 ROOT::Detail::RDF::RLoopManager *fInputLoopManager;
300 bool fToNTuple;
301 bool fIncludeVariations;
302};
303
304template <typename PrevNodeType>
305std::unique_ptr<RActionBase>
306BuildAction(const ColumnNames_t &colNames, const std::shared_ptr<SnapshotHelperArgs> &snapHelperArgs,
307 const unsigned int nSlots, std::shared_ptr<PrevNodeType> prevNode, const RColumnRegister &colRegister,
308 const std::vector<const std::type_info *> &colTypeIDs)
309{
310 const auto &filename = snapHelperArgs->fFileName;
311 const auto &dirname = snapHelperArgs->fDirName;
312 const auto &treename = snapHelperArgs->fTreeName;
313 const auto &outputColNames = snapHelperArgs->fOutputColNames;
314 const auto &options = snapHelperArgs->fOptions;
315 const auto &outputLM = snapHelperArgs->fOutputLoopManager;
316 const auto &inputLM = snapHelperArgs->fInputLoopManager;
317
318 auto sz = colNames.size();
319 std::vector<bool> isDefine(sz);
320 for (auto i = 0u; i < sz; ++i)
321 isDefine[i] = colRegister.IsDefineOrAlias(colNames[i]);
322
323 std::unique_ptr<RActionBase> actionPtr;
324 if (snapHelperArgs->fToNTuple) {
325 // We use the same helper for single- and multi-thread snapshot.
326 using Helper_t = UntypedSnapshotRNTupleHelper;
328
331 colNames, colTypeIDs, prevNode, colRegister));
332 } else {
334 // single-thread snapshot
335 if (snapHelperArgs->fIncludeVariations) {
336 using Helper_t = SnapshotHelperWithVariations;
339 std::move(isDefine), outputLM, inputLM, colTypeIDs),
340 colNames, colTypeIDs, prevNode, colRegister));
341 } else {
342 using Helper_t = UntypedSnapshotTTreeHelper;
345 std::move(isDefine), outputLM, inputLM, colTypeIDs),
346 colNames, colTypeIDs, prevNode, colRegister));
347 }
348 } else {
349 if (snapHelperArgs->fIncludeVariations) {
350 throw std::invalid_argument("Multi-threaded snapshot with variations is not supported yet.");
351 }
352 // multi-thread snapshot
353 using Helper_t = UntypedSnapshotTTreeHelperMT;
356 std::move(isDefine), outputLM, inputLM, colTypeIDs),
357 colNames, colTypeIDs, prevNode, colRegister));
358 }
359 }
360
361 return actionPtr;
362}
363
364// Book with custom helper type
365template <typename... ColTypes, typename PrevNodeType, typename Helper_t>
366std::unique_ptr<RActionBase>
367BuildAction(const ColumnNames_t &bl, const std::shared_ptr<Helper_t> &h, const unsigned int /*nSlots*/,
368 std::shared_ptr<PrevNodeType> prevNode, ActionTags::Book, const RColumnRegister &colRegister)
369{
370 using Action_t = RAction<Helper_t, PrevNodeType, TTraits::TypeList<ColTypes...>>;
371 return std::make_unique<Action_t>(Helper_t(std::move(*h)), bl, std::move(prevNode), colRegister);
372}
373
374/****** end BuildAndBook ******/
375
376template <typename Filter>
377void CheckFilter(Filter &)
378{
379 using FilterRet_t = typename RDF::CallableTraits<Filter>::ret_type;
380 static_assert(std::is_convertible<FilterRet_t, bool>::value,
381 "filter expression returns a type that is not convertible to bool");
382}
383
384ColumnNames_t FilterArraySizeColNames(const ColumnNames_t &columnNames, const std::string &action);
385
386void CheckValidCppVarName(std::string_view var, const std::string &where);
387
388void CheckForRedefinition(const std::string &where, std::string_view definedCol, const RColumnRegister &colRegister,
389 const ColumnNames_t &dataSourceColumns);
390
391void CheckForDefinition(const std::string &where, std::string_view definedColView, const RColumnRegister &colRegister,
392 const ColumnNames_t &dataSourceColumns);
393
394void CheckForNoVariations(const std::string &where, std::string_view definedColView,
395 const RColumnRegister &colRegister);
396
397std::string PrettyPrintAddr(const void *const addr);
398
399std::shared_ptr<RJittedFilter> BookFilterJit(std::shared_ptr<RNodeBase> prevNode, std::string_view name,
400 std::string_view expression, const RColumnRegister &colRegister,
401 TTree *tree, RDataSource *ds);
402
403std::shared_ptr<RJittedDefine> BookDefineJit(std::string_view name, std::string_view expression, RLoopManager &lm,
404 RDataSource *ds, const RColumnRegister &colRegister);
405
406std::shared_ptr<RJittedDefine> BookDefinePerSampleJit(std::string_view name, std::string_view expression,
407 RLoopManager &lm, const RColumnRegister &colRegister);
408
409std::shared_ptr<RJittedVariation>
410BookVariationJit(const std::vector<std::string> &colNames, std::string_view variationName,
411 const std::vector<std::string> &variationTags, std::string_view expression, RLoopManager &lm,
412 RDataSource *ds, const RColumnRegister &colRegister, bool isSingleColumn);
413
414std::string JitBuildAction(const ColumnNames_t &bl, const std::type_info &art, const std::type_info &at, TTree *tree,
415 const unsigned int nSlots, const RColumnRegister &colRegister, RDataSource *ds,
416 const bool vector2RVec = true);
417
418// Allocate a weak_ptr on the heap, return a pointer to it. The user is responsible for deleting this weak_ptr.
419// This function is meant to be used by RInterface's methods that book code for jitting.
420// The problem it solves is that we generate code to be lazily jitted with the addresses of certain objects in them,
421// and we need to check those objects are still alive when the generated code is finally jitted and executed.
422// So we pass addresses to weak_ptrs allocated on the heap to the jitted code, which is then responsible for
423// the deletion of the weak_ptr object.
424template <typename T>
425std::weak_ptr<T> *MakeWeakOnHeap(const std::shared_ptr<T> &shPtr)
426{
427 return new std::weak_ptr<T>(shPtr);
428}
429
430// Same as MakeWeakOnHeap, but create a shared_ptr that makes sure the object is definitely kept alive.
431template <typename T>
432std::shared_ptr<T> *MakeSharedOnHeap(const std::shared_ptr<T> &shPtr)
433{
434 return new std::shared_ptr<T>(shPtr);
435}
436
437bool AtLeastOneEmptyString(const std::vector<std::string_view> strings);
438
439/// Take a shared_ptr<AnyNodeType> and return a shared_ptr<RNodeBase>.
440/// This works for RLoopManager nodes as well as filters and ranges.
441std::shared_ptr<RNodeBase> UpcastNode(std::shared_ptr<RNodeBase> ptr);
442
443ColumnNames_t GetValidatedColumnNames(RLoopManager &lm, const unsigned int nColumns, const ColumnNames_t &columns,
444 const RColumnRegister &validDefines, RDataSource *ds);
445
446std::vector<std::string> GetValidatedArgTypes(const ColumnNames_t &colNames, const RColumnRegister &colRegister,
447 TTree *tree, RDataSource *ds, const std::string &context,
448 bool vector2RVec);
449
450template <typename T>
451void AddDSColumnsHelper(const std::string &colName, RLoopManager &lm, RDataSource &ds, RColumnRegister &colRegister)
452{
453
454 if (colRegister.IsDefineOrAlias(colName))
455 return;
456
457 if (lm.HasDataSourceColumnReaders(colName, typeid(T)))
458 return;
459
460 if (!ds.HasColumn(colName) &&
461 lm.GetSuppressErrorsForMissingBranches().find(colName) == lm.GetSuppressErrorsForMissingBranches().end())
462 return;
463
464 const auto nSlots = lm.GetNSlots();
465 std::vector<std::unique_ptr<RColumnReaderBase>> colReaders;
466 colReaders.reserve(nSlots);
467
468 const auto valuePtrs = ds.GetColumnReaders<T>(colName);
469 if (!valuePtrs.empty()) { // we are using the old GetColumnReaders mechanism in this RDataSource
470 for (auto *ptr : valuePtrs)
471 colReaders.emplace_back(new RDSColumnReader<T>(ptr));
472
473 } else { // using the new GetColumnReaders mechanism
474 // TODO consider changing the interface so we return all of these for all slots in one go
475 for (auto slot = 0u; slot < lm.GetNSlots(); ++slot)
476 colReaders.emplace_back(
477 ROOT::Internal::RDF::CreateColumnReader(ds, slot, colName, typeid(T), /*treeReader*/ nullptr));
478 }
479
480 lm.AddDataSourceColumnReaders(colName, std::move(colReaders), typeid(T));
481}
482
483/// Take list of column names that must be defined, current map of custom columns, current list of defined column names,
484/// and return a new map of custom columns (with the new datasource columns added to it)
485template <typename... ColumnTypes>
486void AddDSColumns(const std::vector<std::string> &requiredCols, RLoopManager &lm, RDataSource &ds,
488{
489 // hack to expand a template parameter pack without c++17 fold expressions.
490 using expander = int[];
491 int i = 0;
493}
494
495void AddDSColumns(const std::vector<std::string> &requiredCols, ROOT::Detail::RDF::RLoopManager &lm,
496 ROOT::RDF::RDataSource &ds, const std::vector<const std::type_info *> &colTypeIDs,
498
499// this function is meant to be called by the jitted code generated by BookFilterJit
500template <typename F>
501void JitFilterHelper(F &&f, const ColumnNames_t &cols, RColumnRegister &colRegister,
503{
504 if (!jittedFilter) {
505 // The branch of the computation graph that needed this jitted code went out of scope between the type
506 // jitting was booked and the time jitting actually happened. Nothing to do other than cleaning up.
507 return;
508 }
509
510 // mock Filter logic -- validity checks and Define-ition of RDataSource columns
511 using Callable_t = std::decay_t<F>;
512 auto prevNode = jittedFilter->MoveOutPrevNode();
513 using PrevNode_t = typename decltype(prevNode)::element_type;
515 using ColTypes_t = typename TTraits::CallableTraits<Callable_t>::arg_types;
516 constexpr auto nColumns = ColTypes_t::list_size;
517 CheckFilter(f);
518
519 auto ds = lm.GetDataSource();
520
521 if (ds != nullptr && !cols.empty())
523
524 jittedFilter->SetFilter(
525 std::unique_ptr<RFilterBase>(new F_t(std::forward<F>(f), cols, prevNode, colRegister, jittedFilter->GetName())));
526}
527
528namespace DefineTypes {
529struct RDefineTag {};
530struct RDefinePerSampleTag {};
531}
532
533template <typename F>
534auto MakeDefineNode(DefineTypes::RDefineTag, std::string_view name, std::string_view dummyType, F &&f,
535 const ColumnNames_t &cols, RColumnRegister &colRegister, RLoopManager &lm)
536{
537 return std::unique_ptr<RDefineBase>(new RDefine<std::decay_t<F>, ExtraArgsForDefine::None>(
538 name, dummyType, std::forward<F>(f), cols, colRegister, lm));
539}
540
541template <typename F>
542auto MakeDefineNode(DefineTypes::RDefinePerSampleTag, std::string_view name, std::string_view dummyType, F &&f,
543 const ColumnNames_t &, RColumnRegister &, RLoopManager &lm)
544{
545 return std::unique_ptr<RDefineBase>(
546 new RDefinePerSample<std::decay_t<F>>(name, dummyType, std::forward<F>(f), lm));
547}
548
549// Build a RDefine or a RDefinePerSample object and attach it to an existing RJittedDefine
550// This function is meant to be called by jitted code right before starting the event loop.
551// If colsPtr is null, build a RDefinePerSample (it has no input columns), otherwise a RDefine.
552template <typename RDefineTypeTag, typename F>
553void JitDefineHelper(F &&f, const ColumnNames_t &cols, RColumnRegister &colRegister,
555{
556
557 if (!jittedDefine) {
558 // The branch of the computation graph that needed this jitted code went out of scope between the type
559 // jitting was booked and the time jitting actually happened. Nothing to do other than cleaning up.
560 return;
561 }
562
563 using Callable_t = std::decay_t<F>;
564 using ColTypes_t = typename TTraits::CallableTraits<Callable_t>::arg_types;
565
566 auto ds = lm.GetDataSource();
567 if (ds != nullptr && !cols.empty())
569
570 // will never actually be used (trumped by jittedDefine->GetTypeName()), but we set it to something meaningful
571 // to help devs debugging
572 const auto dummyType = "jittedCol_t";
573 // use unique_ptr<RDefineBase> instead of make_unique<NewCol_t> to reduce jit/compile-times
574 std::unique_ptr<RDefineBase> newCol{
575 MakeDefineNode(RDefineTypeTag{}, jittedDefine->GetName(), dummyType, std::forward<F>(f), cols, colRegister, lm)};
576 jittedDefine->SetDefine(std::move(newCol));
577}
578
579template <bool IsSingleColumn, typename F>
580void JitVariationHelper(F &&f, const ColumnNames_t &inputColNames, RColumnRegister &colRegister,
582 const ColumnNames_t &variedColNames, const ColumnNames_t &variationTags) noexcept
583{
584
585 if (!jittedVariation) {
586 // The branch of the computation graph that needed this jitted variation went out of scope between the type
587 // jitting was booked and the time jitting actually happened. Nothing to do other than cleaning up.
588 return;
589 }
590
591 using Callable_t = std::decay_t<F>;
592 using ColTypes_t = typename TTraits::CallableTraits<Callable_t>::arg_types;
593
594 auto ds = lm.GetDataSource();
595 if (ds != nullptr && !inputColNames.empty())
597
598 // use unique_ptr<RDefineBase> instead of make_unique<NewCol_t> to reduce jit/compile-times
599 std::unique_ptr<RVariationBase> newVariation{new RVariation<std::decay_t<F>, IsSingleColumn>(
600 variedColNames, jittedVariation->GetVariationName(), std::forward<F>(f), variationTags,
601 jittedVariation->GetTypeName(), colRegister, lm, inputColNames)};
602 jittedVariation->SetVariation(std::move(newVariation));
603}
604
605/// Convenience function invoked by jitted code to build action nodes at runtime
606template <typename ActionTag, typename... ColTypes, typename HelperArgType>
607void CallBuildAction(const ColumnNames_t &cols, RColumnRegister &colRegister, ROOT::Detail::RDF::RLoopManager &lm,
608 RJittedAction *jittedAction, unsigned int nSlots,
609 std::shared_ptr<HelperArgType> *helperArg) noexcept
610{
611 if (!jittedAction) {
612 // The branch of the computation graph that needed this jitted variation went out of scope between the type
613 // jitting was booked and the time jitting actually happened. Nothing to do other than cleaning up.
614 return;
615 }
616
617 using ColTypes_t = TypeList<ColTypes...>;
618 constexpr auto nColumns = ColTypes_t::list_size;
619 auto ds = lm.GetDataSource();
620 if (ds != nullptr && !cols.empty())
622
623 auto actionPtr =
625 jittedAction->SetAction(std::move(actionPtr));
626}
627
628/// The contained `type` alias is `double` if `T == RInferredType`, `U` if `T == std::container<U>`, `T` otherwise.
629template <typename T, bool Container = IsDataContainer<T>::value && !std::is_same<T, std::string>::value>
630struct RMinReturnType {
631 using type = T;
632};
633
634template <>
636 using type = double;
637};
638
639template <typename T>
640struct RMinReturnType<T, true> {
641 using type = TTraits::TakeFirstParameter_t<T>;
642};
643
644// return wrapper around f that prepends an `unsigned int slot` parameter
645template <typename R, typename F, typename... Args>
646std::function<R(unsigned int, Args...)> AddSlotParameter(F &f, TypeList<Args...>)
647{
648 return [f](unsigned int, Args... a) mutable -> R { return f(a...); };
649}
650
651template <typename ColType, typename... Rest>
652struct RNeedJittingHelper {
653 static constexpr bool value = RNeedJittingHelper<Rest...>::value;
654};
655
656template <typename... Rest>
658 static constexpr bool value = true;
659};
660
661template <typename T>
662struct RNeedJittingHelper<T> {
663 static constexpr bool value = false;
664};
665
666template <>
668 static constexpr bool value = true;
669};
670
671template <typename ...ColTypes>
672struct RNeedJitting {
673 static constexpr bool value = RNeedJittingHelper<ColTypes...>::value;
674};
675
676template <>
677struct RNeedJitting<> {
678 static constexpr bool value = false;
679};
680
681///////////////////////////////////////////////////////////////////////////////
682/// Check preconditions for RInterface::Aggregate:
683/// - the aggregator callable must have signature `U(U,T)` or `void(U&,T)`.
684/// - the merge callable must have signature `U(U,U)` or `void(std::vector<U>&)`
685template <typename R, typename Merge, typename U, typename T, typename decayedU = std::decay_t<U>,
686 typename mergeArgsNoDecay_t = typename CallableTraits<Merge>::arg_types_nodecay,
687 typename mergeArgs_t = typename CallableTraits<Merge>::arg_types,
688 typename mergeRet_t = typename CallableTraits<Merge>::ret_type>
690{
691 constexpr bool isAggregatorOk =
692 (std::is_same<R, decayedU>::value) || (std::is_same<R, void>::value && std::is_lvalue_reference<U>::value);
693 static_assert(isAggregatorOk, "aggregator function must have signature `U(U,T)` or `void(U&,T)`");
694 constexpr bool isMergeOk =
695 (std::is_same<TypeList<decayedU, decayedU>, mergeArgs_t>::value && std::is_same<decayedU, mergeRet_t>::value) ||
696 (std::is_same<TypeList<std::vector<decayedU> &>, mergeArgsNoDecay_t>::value &&
697 std::is_same<void, mergeRet_t>::value);
698 static_assert(isMergeOk, "merge function must have signature `U(U,U)` or `void(std::vector<U>&)`");
699}
700
701///////////////////////////////////////////////////////////////////////////////
702/// This overload of CheckAggregate is called when the aggregator takes more than two arguments
703template <typename R, typename T>
704void CheckAggregate(T)
705{
706 static_assert(sizeof(T) == 0, "aggregator function must take exactly two arguments");
707}
708
709///////////////////////////////////////////////////////////////////////////////
710/// Check as many template parameters were passed as the number of column names, throw if this is not the case.
711void CheckTypesAndPars(unsigned int nTemplateParams, unsigned int nColumnNames);
712
713/// Return local BranchNames or default BranchNames according to which one should be used
714const ColumnNames_t SelectColumns(unsigned int nArgs, const ColumnNames_t &bl, const ColumnNames_t &defBl);
715
716/// Check whether column names refer to a valid branch of a TTree or have been `Define`d. Return invalid column names.
717ColumnNames_t FindUnknownColumns(const ColumnNames_t &requiredCols, const RColumnRegister &definedCols,
718 const ColumnNames_t &dataSourceColumns);
719
720/// Returns the list of Filters defined in the whole graph
721std::vector<std::string> GetFilterNames(const std::shared_ptr<RLoopManager> &loopManager);
722
723/// Returns the list of Filters defined in the branch
724template <typename NodeType>
725std::vector<std::string> GetFilterNames(const std::shared_ptr<NodeType> &node)
726{
727 std::vector<std::string> filterNames;
728 node->AddFilterName(filterNames);
729 return filterNames;
730}
731
732struct ParsedTreePath {
733 std::string fTreeName;
734 std::string fDirName;
735};
736
738
739// Check if a condition is true for all types
740template <bool...>
741struct TBoolPack;
742
743template <bool... bs>
744using IsTrueForAllImpl_t = typename std::is_same<TBoolPack<bs..., true>, TBoolPack<true, bs...>>;
745
746template <bool... Conditions>
747struct TEvalAnd {
748 static constexpr bool value = IsTrueForAllImpl_t<Conditions...>::value;
749};
750
751// Check if a class is a specialisation of stl containers templates
752// clang-format off
753
754template <typename>
755struct IsList_t : std::false_type {};
756
757template <typename T>
758struct IsList_t<std::list<T>> : std::true_type {};
759
760template <typename>
761struct IsDeque_t : std::false_type {};
762
763template <typename T>
764struct IsDeque_t<std::deque<T>> : std::true_type {};
765// clang-format on
766
767void CheckForDuplicateSnapshotColumns(const ColumnNames_t &cols);
768
770
771template <typename T>
772struct InnerValueType {
773 using type = T; // fallback for when T is not a nested RVec
774};
775
776template <typename Elem>
777struct InnerValueType<ROOT::VecOps::RVec<ROOT::VecOps::RVec<Elem>>> {
778 using type = Elem;
779};
780
781template <typename T>
783
784std::pair<std::vector<std::string>, std::vector<std::string>>
786 std::vector<std::string> &&colsWithAliases);
787
788void RemoveDuplicates(ColumnNames_t &columnNames);
789void RemoveRNTupleSubFields(ColumnNames_t &columnNames);
790
791} // namespace RDF
792} // namespace Internal
793
794namespace Detail {
795namespace RDF {
796
797/// The aliased type is `double` if `T == RInferredType`, `U` if `T == container<U>`, `T` otherwise.
798template <typename T>
799using MinReturnType_t = typename RDFInternal::RMinReturnType<T>::type;
800
801template <typename T>
803
804template <typename T>
806
807} // namespace RDF
808} // namespace Detail
809} // namespace ROOT
810
811/// \endcond
812
813#endif
#define f(i)
Definition RSha256.hxx:104
#define g(i)
Definition RSha256.hxx:105
#define a(i)
Definition RSha256.hxx:99
#define h(i)
Definition RSha256.hxx:106
ROOT::Detail::TRangeCast< T, true > TRangeDynCast
TRangeDynCast is an adapter class that allows the typed iteration through a TCollection.
Int_t gErrorIgnoreLevel
errors with level below this value will be ignored. Default is kUnset.
Definition TError.cxx:33
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char filename
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void value
Option_t Option_t TPoint TPoint const char GetTextMagnitude GetFillStyle GetLineColor GetLineWidth GetMarkerStyle GetTextAlign GetTextColor GetTextSize void char Point_t Rectangle_t WindowAttributes_t Float_t Float_t Float_t Int_t Int_t UInt_t UInt_t Rectangle_t Int_t Int_t Window_t TString Int_t GCValues_t GetPrimarySelectionOwner GetDisplay GetScreen GetColormap GetNativeEvent const char const char dpyName wid window const char font_name cursor keysym reg const char only_if_exist regb h Point_t winding char text const char depth char const char Int_t count const char ColorStruct_t color const char Pixmap_t Pixmap_t PictureAttributes_t attr const char char ret_data h unsigned char height h Atom_t Int_t ULong_t ULong_t unsigned char prop_list Atom_t Atom_t Atom_t Time_t type
char name[80]
Definition TGX11.cxx:110
A wrapper around a concrete RDefine, which forwards all calls to it RJittedDefine is a placeholder th...
A wrapper around a concrete RFilter, which forwards all calls to it RJittedFilter is the type of the ...
The head node of a RDF computation graph.
A histogram for aggregation of data along multiple dimensions.
Definition RHist.hxx:64
A binder for user-defined columns, variations and aliases.
RDataSource defines an API that RDataFrame can use to read arbitrary data formats.
The public interface to the RDataFrame federation of classes.
const_iterator end() const
A "std::vector"-like collection of values implementing handy operation to analyse them.
Definition RVec.hxx:1524
A TTree represents a columnar dataset.
Definition TTree.h:89
#define F(x, y, z)
const ColumnNames_t SelectColumns(unsigned int nRequiredNames, const ColumnNames_t &names, const ColumnNames_t &defaultNames)
Choose between local column names or default column names, throw in case of errors.
void CheckForNoVariations(const std::string &where, std::string_view definedColView, const RColumnRegister &colRegister)
Throw if the column has systematic variations attached.
ParsedTreePath ParseTreePath(std::string_view fullTreeName)
std::shared_ptr< RJittedDefine > BookDefinePerSampleJit(std::string_view name, std::string_view expression, RLoopManager &lm, const RColumnRegister &colRegister)
Book the jitting of a DefinePerSample call.
void CheckValidCppVarName(std::string_view var, const std::string &where)
void RemoveDuplicates(ColumnNames_t &columnNames)
ColumnNames_t GetValidatedColumnNames(RLoopManager &lm, const unsigned int nColumns, const ColumnNames_t &columns, const RColumnRegister &colRegister, RDataSource *ds)
Given the desired number of columns and the user-provided list of columns:
std::shared_ptr< RNodeBase > UpcastNode(std::shared_ptr< RNodeBase > ptr)
void CheckSnapshotOptionsFormatCompatibility(const ROOT::RDF::RSnapshotOptions &opts)
void CheckForDefinition(const std::string &where, std::string_view definedColView, const RColumnRegister &colRegister, const ColumnNames_t &dataSourceColumns)
Throw if column definedColView is not already there.
std::vector< std::string > GetFilterNames(const std::shared_ptr< RLoopManager > &loopManager)
std::string PrettyPrintAddr(const void *const addr)
std::shared_ptr< RDFDetail::RJittedFilter > BookFilterJit(std::shared_ptr< RDFDetail::RNodeBase > prevNode, std::string_view name, std::string_view expression, const RColumnRegister &colRegister, TTree *tree, RDataSource *ds)
Book the jitting of a Filter call.
std::string JitBuildAction(const ColumnNames_t &cols, const std::type_info &helperArgType, const std::type_info &at, TTree *tree, const unsigned int nSlots, const RColumnRegister &colRegister, RDataSource *ds, const bool vector2RVec)
void CheckTypesAndPars(unsigned int nTemplateParams, unsigned int nColumnNames)
std::string DemangleTypeIdName(const std::type_info &typeInfo)
bool AtLeastOneEmptyString(const std::vector< std::string_view > strings)
std::unique_ptr< ROOT::Detail::RDF::RColumnReaderBase > CreateColumnReader(ROOT::RDF::RDataSource &ds, unsigned int slot, std::string_view col, const std::type_info &tid, TTreeReader *treeReader)
Definition RDFUtils.cxx:684
std::pair< std::vector< std::string >, std::vector< std::string > > AddSizeBranches(ROOT::RDF::RDataSource *ds, std::vector< std::string > &&colsWithoutAliases, std::vector< std::string > &&colsWithAliases)
Return copies of colsWithoutAliases and colsWithAliases with size branches for variable-sized array b...
void RemoveRNTupleSubFields(ColumnNames_t &columnNames)
ColumnNames_t FilterArraySizeColNames(const ColumnNames_t &columnNames, const std::string &action)
Take a list of column names, return that list with entries starting by '#' filtered out.
std::shared_ptr< RJittedVariation > BookVariationJit(const std::vector< std::string > &colNames, std::string_view variationName, const std::vector< std::string > &variationTags, std::string_view expression, RLoopManager &lm, RDataSource *ds, const RColumnRegister &colRegister, bool isSingleColumn)
Book the jitting of a Vary call.
std::vector< std::string > GetValidatedArgTypes(const ColumnNames_t &colNames, const RColumnRegister &colRegister, TTree *tree, RDataSource *ds, const std::string &context, bool vector2RVec)
void CheckForDuplicateSnapshotColumns(const ColumnNames_t &cols)
ColumnNames_t ConvertRegexToColumns(const ColumnNames_t &colNames, std::string_view columnNameRegexp, std::string_view callerName)
ColumnNames_t FindUnknownColumns(const ColumnNames_t &requiredCols, const RColumnRegister &definedCols, const ColumnNames_t &dataSourceColumns)
unsigned int & NThreadPerTH3()
Obtain or set the number of threads that will share a clone of a thread-safe 3D histogram.
Definition RDFUtils.cxx:76
void CheckForRedefinition(const std::string &where, std::string_view definedColView, const RColumnRegister &colRegister, const ColumnNames_t &dataSourceColumns)
Throw if column definedColView is already there.
std::shared_ptr< RJittedDefine > BookDefineJit(std::string_view name, std::string_view expression, RLoopManager &lm, RDataSource *ds, const RColumnRegister &colRegister)
Book the jitting of a Define call.
ROOT type_traits extensions.
Bool_t IsImplicitMTEnabled()
Returns true if the implicit multi-threading in ROOT is enabled.
Definition TROOT.cxx:600
A collection of options to steer the creation of the dataset on disk through Snapshot().
Lightweight storage for a collection of types.